Everyone Expects Two AI Blocs. The Gulf Is Betting It Doesn't Have to Choose

The conventional framing of the global AI race splits cleanly into two camps: the United States, with Nvidia, the hyperscalers and most leading frontier labs, on one side, and China, with its own models, engineering capacity and increasingly competitive stack, on the other. A new analysis from Judah Taub, founder of Israeli venture firm Hetz Ventures, published in Semafor, argues that Saudi Arabia and the UAE are deliberately refusing to pick a side, and that the refusal itself is the strategy, not an oversight or a diplomatic hedge waiting to collapse under pressure.

The evidence sits in plain view. Saudi Arabia has become one of the deepest American AI partners in the Gulf, with AWS alone committing $5.3 billion to data centres in the Kingdom. Yet at Riyadh's LEAP conference earlier this month, Saudi state-backed HUMAIN unveiled a flagship Arabic-language frontier model built on architecture from China's MiniMax, a 428-billion-parameter system that topped seven Arabic-language benchmarks against competing models from OpenAI and Anthropic. Rather than treating these as contradictory signals from a country still deciding which patron to choose, Taub argues they are two deliberate legs of the same strategy: extracting the best available technology from whichever ecosystem currently leads in a given category, rather than importing one stack wholesale and foreclosing the other.

The framing that makes this more than opportunism is capability, not just appetite. Saudi Arabia and the UAE aren't attempting to replicate either the American or Chinese AI stack end to end, model architecture, chip design, cloud infrastructure and developer ecosystem all at once. Taub's argument is that they don't need to: their genuine strategic value lies in being among the very few places on earth where both ecosystems can operate simultaneously, each treating the Gulf as a market too significant to cede to the other. Washington increasingly treats advanced chip access as a form of foreign policy, comparable to controlling sensitive military technology rather than a commercial transaction, while Beijing sees open models and an increasingly capable domestic stack as its own pathway into markets its hardware alone cannot reach. Riyadh's calculus, in Taub's framing, is simpler than either capital's: not having to choose between them.

This reading gives sharper geopolitical context to a story this site covered directly, the formal US-UAE AI Acceleration Partnership and the parallel chip-access, localisation and manufacturing track running through HUMAIN specifically. Read in isolation, that bilateral partnership looked like a straightforward alignment toward Washington. Read alongside HUMAIN's Chinese-architecture Arabic model, it looks more like one carefully managed leg of a two-legged position, with the American partnership securing chip access and security-standard credibility while the Chinese-model partnership secures technical capability and independence from any single supplier's roadmap or export-control regime.

The strategy stems from a simple calculation. American restrictions push Riyadh toward Chinese alternatives, and Chinese shortfalls in frontier performance push it back toward Washington. In the end, both sides end up competing for something the Gulf increasingly controls: AI infrastructure, the same compute capacity race already visible in the UAE's own push to become a producer, not just a consumer, of frontier AI hardware.

The historical parallel Taub draws is a deliberate one: rather than two sealed blocs competing for exclusive territory, the AI race may increasingly resemble a Cold War dynamic with a handful of swing states positioned to play both sides, extracting economic and strategic benefit from each without being absorbed into either. The Gulf, Taub notes, spent decades converting control over energy into geopolitical leverage; it is now attempting the same playbook with compute, treating capital, energy and land, its genuinely scarce and desirable assets in an AI-constrained world, as the currency that buys access to both ecosystems rather than forcing a binary choice between them.

For enterprises and investors evaluating Gulf AI partnerships specifically, the practical implication is that dual-sourcing at the state level is likely to become a durable feature of the region's AI landscape rather than a transitional phase before consolidation around one bloc. A company or investor expecting Saudi Arabia or the UAE to eventually "pick a side" as chip export politics intensify may be waiting for a resolution the strategy is specifically designed never to reach.